Jim, Re: "Trust me, they fixed the benchmarks. Especially since Intel stated their goal to find benchmarks that make P4 look good."
The idea, from Intel's perspective, is to prove that newer applications and usage models take more advantage of the hardware than legacy applications, and this is also in line with what they have stated. It's clearly stated in the SysMark 2001 benchmark documentation, that newer, richer application mixes were used to more closely represent real world situations. Check out the description of the Usage Model here. You will see that it differs from the Usage Model in SysMark 2000, which explains why the scores vary a lot between the two versions.
bapco.com
"Fixing" the benchmark suggests some kind of Dan3 Conspiracy Theory that Intel intentionally crippled the competitors scores, just so that they could look better. If you read the whitepaper, though, you will see that Bapco uses a different way of testing that other productivity packages lack. Not everyone will agree that their method of testing is the right one, but that's why multiple benchmarks exist, giving the reader a chance to see several data points. You should realize that there is a difference between promoting your product, and dissing the competition. Publicly, Intel does the former, while AMD (or at least Jerry Sanders) does the latter.
wanna_bmw |
